AI Denied Your Mortgage — Before You Even Applied 😳

AI Denied Your Mortgage — Before You Even Applied 😳

May 29, 20253 min read

Imagine getting denied for a mortgage… before you even apply. Not by a person. Not by a bank. But by an algorithm — a machine trained on massive amounts of historical data, making decisions in milliseconds while you’re casually browsing homes online.

This isn’t fiction. It’s today’s reality. And it's happening to families like Marcus and Kesha.


A Story You Haven’t Heard — But Should

Marcus works maintenance for a school district. Kesha’s a nurse. They’ve been married 12 years, rent a small duplex in San Bernardino, and have two kids: Jaylen (8) and Zoe (5). For three years, they saved every tax refund and every overtime shift — $47,000 in the bank, credit scores of 720 and 698, not a single late payment.

One Sunday night, they find a 3-bedroom home in Fontana listed for $485,000. Zillow says they’re good. The lender’s pre-qual tool says likely approved. They fill out the application, upload all their docs — pay stubs, W-2s, bank statements — and hit submit.

Eighteen seconds later: denied.
No phone call. Just a generic email.

They call to ask why. After some digging, they’re told the “automated underwriting system” flagged them. No clear reason, just a vague suggestion to check their credit. But their credit was clean.

They try a second lender. Same result.


The Real Reason Behind the Denials

Here’s what really happened.

The AI didn’t just look at their credit. It scanned everything:

  • Zip code: San Bernardino, flagged as high-risk due to the 2008 crash.

  • Employment gap: Marcus had a two-month break in 2020, flagged as income volatility.

  • Rental address: Previous tenants at the duplex had defaulted on car loans, so the address itself was tagged.

The system didn’t see responsible parents with savings and steady jobs. It saw risk markers.

This is called algorithmic profiling — and it’s changing lending as we know it.


The Quiet Power of Automated Valuations

As if that weren’t enough, their landlord decided to sell the duplex. When Marcus and Kesha checked the home’s value online, it had dropped $73,000 in just six months. Same house. Same neighborhood. No visible changes — except one: the algorithm’s opinion.

According to a 2023 investigation by Business Insider, large hedge fund-owned homes often see inflated valuations on platforms like Zillow, while homes in working-class areas see their values quietly drop:
https://www.businessinsider.com/zillow-home-values-real-estate-hedge-fund-bias-2023-8

And it’s not just Zillow. Appraisals are going AI too — using tools called AVMs (Automated Valuation Models). Supposed to be neutral, these tools often undervalue homes in Black and Latino neighborhoods by $15,000–$30,000, according to a 2022 Brookings study:
https://www.brookings.edu/articles/devaluation-of-assets-in-black-neighborhoods/

Why? Because the data was trained on decades of redlining and discrimination. The algorithm learns bias — and then calls it “fair.”


Even a Perfect File Can Be Rejected

Marcus and Kesha eventually find a credit union that still uses human underwriters. The appraiser visits the Fontana home and values it right where expected: $485,000.

But the AVM?

It said $441,000.

That $44,000 gap killed the loan — even though the human said yes.

That’s not just a tech issue. That’s digital redlining.


Here’s What You Need to Know

When an AI denies your loan, there’s often:

  • No explanation

  • No appeal

  • No second chance

Most loan officers don’t even understand how these systems work. They just input your data and read the result.

But here’s the truth:
Every system has a workaround.
Every "no" can be challenged — if you know how.


How to Fight Back

You’re not crazy. You’re not alone. And you’re not stuck.

After nearly 30 years in the industry, I’ve seen how this system thinks — and more importantly, how to beat it. I can’t guarantee approval, but I can promise you this: you’ll finally understand what’s happening and walk away with a real plan, not just false hope.

Because at the end of the day, algorithms don’t buy homes. People do.
And every family deserves a shot — even if they have to fight a robot to get there.

Back to Blog
loanDepot Ramin Ekhtiar
The High Desert Group Logo

Social Media Links

Contact Us

(408) 431-7296

17500 Depot St, Ste 17 Morgan Hill, CA 95037

Copyright 2025 . All rights reserved. Ramin Ekhtiar NMLS #1469025 loanDepot Company NMLS# 174457 | Equal Housing Opportunity | Equal Housing Lender